https://t.co/sLwHkxbS7u might be the best tool i've used with an agent.
i told my agent what i wanted through @pendev's mcp. no style guide, no references, no design instructions.
it made this.
opus 5.5 alone has never given me a landing page this good with so little direction.
this sent me down a rabbit hole.
inspired by @NielsRoggeโs explanation and @harshagundalโs demo, i built Decision Lab: a 350M model that makes structured decisions locally in your browser with WebGPU.
83% on 1,600 held-out examples.
open source: https://t.co/PiNfflv6g0
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
Iโve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
โข 20-200x faster
โข 40-400x cheaper (w/ output tokens free)
โข Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution